Time-varying covariates and coefficients in Cox regression models
|
|
- Gertrude Houston
- 5 years ago
- Views:
Transcription
1 Big-data Clinical Trial Column Page 1 of 10 -varying covariates and coefficients in Cox regression models Zhongheng Zhang 1, Jaakko Reinikainen 2, Kazeem Adedayo Adeleke 3, Marcel E. Pieterse 4, Catharina G. M. Groothuis-Oudshoorn 5 1 Department of Emergency Medicine, Sir Run-Run Shaw Hospital, Zhejiang University School of Medicine, Hangzhou , China; 2 Department of Public Health Solutions, National Institute for Health and Welfare, Helsinki, Finland; 3 Department of Mathematics, Obafemi Awolowo University, Nigeria; 4 Department of Psychology, Health and Technology, Centre for ehealth & Well-being Research, 5 Department of Health Technology and Services Research, University of Twente, Enschede, The Netherlands Correspondence to: Zhongheng Zhang. No. 3, East Qingchun Road, Hangzhou , China. zh_zhang1984@zju.edu.cn. Abstract: -varying covariance occurs when a covariate changes over time during the follow-up period. Such variable can be analyzed with the Cox regression model to estimate its effect on survival time. For this it is essential to organize the data in a counting process style. In situations when the proportional hazards assumption of the Cox regression model does not hold, we say that the effect of the covariate is time-varying. The proportional hazards assumption can be tested by examining the residuals of the model. The rejection of the null hypothesis induces the use of time varying coefficient to describe the data. The time varying coefficient can be described with a step function or a parametric time function. This article aims to illustrate how to carry out statistical analyses in the presence of time-varying covariates or coefficients with R. Keywords: Cox proportional hazards; time dependent; time varying; Schoenfeld residuals; time-to-event Submitted Dec 05, Accepted for publication Jan 04, doi: /atm View this article at: Introduction -varying covariance occurs when a given covariate changes over time during the follow-up period, which is a common phenomenon in clinical research. For example, in a patient with sepsis, the C-reactive protein (CRP) may be measured repeatedly to evaluate inflammatory status until it returns normal (1). In clinical oncology, the recurrence status of a patient is usually checked at a predefined time interval. In many cases when studying the relation between a survival outcome and covariate(s), investigators will only consider the baseline value of the covariate, which however, fails to consider the relation of the survival outcome as a function of the change of the covariate. For example, the effect of smoking on cancer risk has been extensively studied. However, the smoking status is ever changing during the follow up period (2). Such a covariate can be considered as a time-varying covariate. -varying covariates can be classified as either internal, when the path is affected by survival status, or external, when the covariate is the fixed/defined covariate (3). An internal covariate is typically the output of a stochastic process generated by an individual under study and observed only as long as the subject survives and uncensored. Thus, such data are found in clinical trials where records of patients general condition are made at regular intervals and where failure time occurs when the patient dies. An external covariate X( ), in contrast, may influence the rate of failure over time, but its path up to time t > v is not affected by the occurrence of failure time at time v. It is also a derived or predetermined covariate. Examples of external covariates are age of an individual in a trial of long duration, or a measure of airborne pollution as a predictor of the frequency of asthma attacks. The covariate allows incorporation of a time interaction function X(t) or X(g(t)) (4,5). Consider the general hazard model for failure time proposed by Cox [1972] (6),
2 Page 2 of 10 Zhang et al. -varying covariate and coefficient in survival analysis λ(t X) = λ 0 (t)exp (β' X) [1] where λ 0 (t) is the baseline hazard function (possibly nondistributional) and β' = (β 1, β 2,..., β p ) is a vector of regression coefficients. In the simple form of the Cox model, X is a vector of time-fixed covariates. One approach for using time-varying covariate data is to extend the Cox proportional hazard model to allow timevarying covariates (7). λ(t Z(t)) = λ 0 (t)exp (β'x + γ'xg(t)) [2] where β' and γ' are coefficients of time-fixed and timevarying covariate respectively. Suppose we let Z(t) represent the covariate, then Z(t) = [x 1, x 2... x p, X 1 g(t), X 2 g(t)..., X q g(t)] [3] and the hazard ratio is λ(; tzt () = ( ) = exp ( ' + ' ( ) ) [4] λ(; tzt () * * HR β x γ Xg t * which is a non-constant hazard rate. Such functionality can be implemented in many sophisticated software and here we will illustrate how to perform such kind of analysis with R-program (8). The main approaches for survival analysis with time-varying covariates are time-dependent Cox models (7) and the joint modeling of longitudinal and survival data (9). -dependent Cox models are more appropriate for external covariates (e.g., external covariates vary as a function of time, independent of the failure time) and are considered in this paper. In a slightly comparable situation, a covariate is measured at baseline but its effect on the outcome is not constant over the follow-up time, which is a violation of the proportional hazards assumption (7). In that case a time-varying coefficient can be incorporated into the Cox regression model to fit such kind of data. In fact, to check the proportional hazards assumption after fitting a Cox regression model is the same as identifying time-varying coefficients. In this paper, we will also show how to check the proportional hazards assumption after fitting a Cox regression model, and in case there is a violation to the assumption, show how the model should be modified to best describe the data. In fact, if the time-varying coefficient can be written as g(β,t) = βg(t), the model with a time-varying coefficient can be expressed as a model with time-varying covariate with a constant coefficient (10). The hazard of failure is related to the covariate by the equation: λ(t X) = λ 0 (t) exp {g(β, t)x} where β is a coefficient related to the covariate X and g(β, t) is a specific function of time that can be defined by investigators. If g(β, t) is a simple function, it can be written as g(β, t) = βg(t). Then the hazard function can be written as: λ(t X) = λ 0 (t) exp {βg,(t)x} = λ 0 (t) exp {βx,(t)} where X(t)= g(t)x. This equation shows that a time-varying coefficient (g(β, t)) model can be modelled with a set of time-varying covariates (X(t)) (10). Verification of proportionality assumption can be done by any of the two known approaches, graphical visualization and numerical approaches. Later in this work we look at Schoenfeld residual scaled plot and log(-log(s(t))) plot. For each predictor in the model, Schoenfeld residual are defined and the residuals for the predictors are plotted against the ranked/transformed failure time. Working example on time-varying covariates To show how to estimate a survival model with timevarying covariates we will construct a simulated dataset. To show how to combine such data we will therefore simulate two data frames in R, one containing the baseline covariates (age and group) and the other a time-varying covariate. With the package survsim (11), a dataset of 100 patients involving continuous and categorical covariates, and a time-to-event outcome can be generated. The simulated dataset is for illustration purpose only and there is no clinical relevance. > library(survsim) > N=100 #number of patients > set.seed(123) > df.tf<-simple.surv.sim(#baseline time fixed n=n, foltime=500, dist.ev=c('llogistic'), anc.ev=c(0.68), beta0.ev=c(5.8), anc.cens=1.2, beta0.cens=7.4, z=list(c("unif", 0.8, 1.2)), beta=list(c(-0.4),c(0)), x=list(c("bern", 0.5), c("normal", 70, 13))) > names(df.tf)[c(1,6,7)]<-c("id","grp","age")
3 Annals of Translational Medicine, Vol 6, No 7 April 2018 Page 3 of 10 The above code generates a data frame containing two time-fixed variables named grp (abbreviated from group) and age. The age variable is assumed to be normally distributed with the mean=70 and standard deviation of 13. The grp variable is a factor (categorical or binary) variable with two levels 0 and 1. The status variable is the outcome status at the corresponding time point. The start and stop variables define the start and stop time points of a followup interval for each individual. The underlying mechanisms of the data generation is beyond the scope of this paper, but interested readers can consult the R document by typing?simple.surv.sim. Alternatively, survival times with timevarying covariates can be generated following the methods proposed by Austin (12). Next, we generate a data frame in a counting process data structure (13), in which each individual is represented by one or more rows. In such a data frame, each row represents a follow up time interval at which the value of a covariate is recorded. The following code generates a timevarying covariate named crp (C-reactive protein) which is assumed to have a normal distribution with a mean of 100 and a standard deviation of 40. In reality, the crp value may be skewed. Although a simple binary covariate such as transplantation, surgery or starting of medication, could be good for demonstrating, we incorporate a numeric variable because such kind of variables are common in reality. > set.seed(123) > nft<-sample(1:10, N,replace=T)#number of follow up time points crp<-round(abs(rnorm(sum(nft)+n, mean=100,sd=40)),1) time<-na id<-na i=0 for(n in nft){ i=i+1 time.n<-sample(1:500,n) time.n<-c(0,sort(time.n)) time<-c(time,time.n) id.n<-rep(i,n+1) id<-c(id,id.n) } df.td <- cbind(data.frame(id,time)[-1,],crp) The number of follow up time intervals is randomly generated for each subject with a maximum of 10. With a for loop function, crp values are assigned to each follow up time interval. A variable id is generated in the for loop, which is a tag for identification of a distinct subject. Finally, a data frame named df.td containing a time-varying covariate crp is generated. Merging data frames with tmerge() function Data containing information on time-varying covariates is often stored in different format than what is required by statistical programs. The first step in analyzing timevarying covariates in survival analysis is to reshape the data frame so that there are multiple rows (time intervals) for each subject, along with covariate values that apply across these intervals. Such a format is also known as the counting process style or (start, stop) form of data. The survival package provides a good function tmerge() for this purpose (7,14). The function usually runs in multiple passes, with the first run defining the basic structure and subsequent runs add variables to that structure. This run does not change the values of original variables but it defines the basic structure of the df object, which is essential for subsequent steps. > df<-tmerge(df.tf,df.tf,id=id, endpt=event(stop,status)) > head(round(df)) id status start stop z grp age tstart tstop endpt In the new data frame df, several variables are added including tstart and tstop representing the start and stop of the follow up interval. The variable endpt is the same as the variable status indicating whether the event is observed. The variable z is the individual heterogeneity generated in the simple.surv.sim() function according to the specified distribution. The value of z is 1 for most observations because the use of the round() function. Next, the total follow-up time is split into the simulated time intervals from dataframe df.td and these intervals are censored by the total follow-up time.
4 Page 4 of 10 Zhang et al. -varying covariate and coefficient in survival analysis > df <- tmerge(df,df.td,id=id, crp=tdc(time,crp)) > head(round(df),10) id status start stop z grp age tstart tstop endpt crp The first argument of tmerge() function is the primary dataset to which new covariates will be added. The second argument is another dataset that contains new covariates. The newname=tdc(y,x) argument creates a new timevarying covariate. The argument y is on the scale of start and stop time. The second argument x is not mandatory. If x is missing the count variable starts at 0 for each subject and becomes 1 at the time of the event. In case x is present the count is set to the value of x. In the example, a crp value is added at each interval. The updated data frame df is in the counting process style that each subject can take several rows. For example, subject 2 has 6 rows, and crp values are different at each row. However, the values of time-fixed variables such as age and grp are consistent within each subject. Fitting the Cox model with a time-varying covariate Next, we will model the survival times as a function of group, age and crp values with Cox regression: > fit.tdc <- coxph(surv(tstart,tstop,endpt)~ grp+age+crp+cluster(id),df) > fit.tdc Call: coxph(formula = Surv(tstart, tstop, endpt) ~ grp + age + crp +, data = df) coef exp(coef) se(coef) robust se z p grp age crp Likelihood ratio test=8.21 on 3 df, p= n= 365, number of events= 67 With the reshaped dataset, the fitting of Cox regression model is straightforward. The output of the coxph() function shows that there is only one hazard ratio (exp(coef)) for the variable crp, which is similar for the two time-fixed covariates age and grp. In the Cox regression model with time-varying covariates, the follow-up time of each subject is divided into shorter time intervals. However, we do not have to take into account in the analysis that individuals may have multiple rows unless there are multiple events per individual. The likelihood equations use information on only at most one row per an individual at any time point, since the time intervals of an individual do not overlap (7). -varying coefficients As noted above, time-varying effect emerges when the proportional hazards assumption is not fulfilled. So, to identify time-varying coefficients is actually to test the proportional hazards assumption after fitting a Cox proportional hazard model. The examination of proportional hazards assumption can be performed using the cox.zph() function shipped with the survival package (14). Below, the lung dataset available from the package survival is employed to illustrate how to explore the proportional assumption. > fit2 <- coxph(surv(time, status) ~ age +ph.karno+sex, data=lung) > zph <- cox.zph(fit2) > zph rho chisq p age ph.karno sex GLOBAL NA First, a Cox proportional model is fit by using the coxph() function. The left side of the formula is the response variable defined by the Surv() function with the followup time and status for each patient. The right side displays
5 Annals of Translational Medicine, Vol 6, No 7 April 2018 Page 5 of 10 Beta(t) for ph.karno Reference line for null effect Average hazard over time -varying hazardv Figure 1 The effect of the covariate ph.karno on mortality outcome varies over time. The horizontal time axis is in km transformed scale, which is the default setting in the cox.zph() function. The dashed lines are lower and upper limits of confidence interval of the effect of ph.karno. It is noted that the effect of ph.karno is not time constant. HR1 HR2 HR3 6 months 2 m HR4 5 months 6 months Figure 2 stratified effect of fixed baseline covariate on survival. Note that the effects of baseline covariate for different time windows are different, resulting in a series of hazard ratios. covariates age (in years), ph.karno (Karnofsky performance score rated by physician) and sex (male =1; female =2). The data argument defines the data frame that contains the data with the variables from the formula. The cox.zph() function is the core to the investigation of proportional hazards assumption. The first argument of the function is an object returned by coxph() function. The output of the function is a table in matrix format with each row representing one variable and the last row is the Schoenfeld s global test for the violation of proportional assumption (15). Columns of the matrix from left to right show the correlation coefficient between transformed survival time and the scaled Schoenfeld residuals (rho), a chi-square statistic (chisq), and the two-sided P value (p). There is no appropriate correlation for the global test, so an NA is entered in the rho column. The result shows that there is significant deviation from the proportional hazards assumption for the variable ph.karno (P= ). The result can be visualized with generic plot() function. In general, an associated global significant test gives a P value ( ) which is an indication of lack of fit of the model. > plot(zph[2],lwd=2) > abline(0,0, col=1,lty=3,lwd=2) > abline(h= fit2$coef[2], col=3, lwd=2, lty=2) > legend("bottomright", legend=c('reference line for null effect', "Average hazard over time", "-varying hazard"), lty=c(3,2,1), col=c(1,3,1), lwd=2) Figure 1 shows the time-varying coefficient for the variable ph.karno. Note that the time axis is not in linear scale because we used km transformation for the time. So in this example we have identified a time-varying coefficients as there appears to be two turning points approximately at values of 180 (the point where the slope of the beta reverses) and 350 (the point where the hazard of the coefficient exceeds the reference for null effect), at which the analysis time can be divided. Step function to explore time-varying coefficient One way to model time-varying coefficients is to use a step function, e.g., (g(t) = I(t t o )), where t o is a specified value. The idea of this method is to split the analysis time into several intervals and Cox proportional model is stratified for these time intervals. The effect of fixed baseline covariates becomes stronger or weaker over time, which can be explored via stratification by time. As illustrated in Figure 2, the effect of the baseline risk factor ph_karno varies over time, resulting in a series of HRs. With the survsplit() function one can split each record into subrecords at prespecified cut time points in the counting process style as we have seen before. > lung.split <- survsplit(surv(time, status) ~., data= lung, cut=c(180, 350),
6 Page 6 of 10 Zhang et al. -varying covariate and coefficient in survival analysis episode= "tgroup", id="id") > head(lung.split[-c(1,4,6:8)]) age sex ph.karno id tstart time status tgroup The first argument of the survsplit() function is a model formula where the model of the survival data can be specified as we have seen before. The cut argument is a vector of cutoff time points. In the example, we cut the analysis time at 180 and 350. The episode option defines a new variable name that will appear in the new data frame. Here, it is tgroup. The resulting data frame is in a counting process form so that each subject is split and takes several rows. For example, patient 2 takes three rows. The original row of (0, 455] with a cut vector of (180, 350) will be split into intervals of (0, 180], (180, 350] and (350, 455]. The newly defined variable tgroup identifies which interval each row belongs to. To explain, tgroup=1 identifies the first time interval (0, 180], and tgroup=2 identifies the second time interval (180, 350]. > fit.split <- coxph(surv(tstart, time, status) ~ age + ph.karno:strata(tgroup)+ sex, data=lung.split) > fit.split Call: coxph(formula = Surv(tstart, time, status) ~ age + ph.karno:strata(tgroup) + sex, data = lung.split) coef exp(coef) se(coef) z p age sex ph.karno:strata(tgroup) tgroup=1 ph.karno:strata(tgroup) tgroup=2 ph.karno:strata(tgroup) tgroup= Likelihood ratio test=27.2 on 5 df, p=5.21e-05 n= 460, number of events= 164 (1 observation deleted due to missingness) Now the Cox regression model is fit as usual, except that it is stratified by the tgroup variable. From the output, it appears that the variable ph.karno only has a significant effect in the first time interval (tgroup=1). The corresponding HR was 0.97 (P= ). The effects of ph.karno on the remaining two time-windows are not statistically significant. Next, we can take a look at the proportional hazards assumption of this stratified Cox regression model. > cox.zph(fit.split) rho chisq p age sex ph.karno:strata(tgroup) tgroup=1 ph.karno:strata(tgroup) tgroup=2 ph.karno:strata(tgroup) tgroup=3 GLOBAL NA The result shows now that there is no correlation between transformed survival time and the scaled Schoenfeld residuals, indicating that the proportional hazards assumption is not violated with the stratified analysis, and judging by the global p-value, the model is fit. Continuous function to describe the time-varying coefficient An alternative method to describe the time-varying coefficient is with a parametric continuous function that is specified by the user. Here we illustrate how to perform such an analysis. >fit.tt <- coxph(surv(time, status) ~ age + ph.karno + tt(ph.karno)+ sex, data=lung, tt = function(x, t,...) x * log(t+20)) In the coxph() function, there is a tt argument to specify the specific transformation of time. In our example, the tt
7 Annals of Translational Medicine, Vol 6, No 7 April 2018 Page 7 of survival by using the abline() function. Beta(t) for ph.karno Figure 3 A parametric time function is assigned to ph.karno. If the time axis is transformed by the function log(t+20), the effect is linear with the slope of (red line). function is defined as tt = function(x, t,...) x * log(t+20), where x is a fixed covariate with time-varying effect, and t is the analysis time. The tt() function is applied to the variable ph.karno in the model formula as tt(ph.karno). The outcomes of the analysis are: > fit.tt Call: coxph(formula = Surv(time, status) ~ age + ph.karno + tt(ph. karno) + sex, data = lung, tt = function(x, t,...) x * log(t + 20)) coef exp(coef) se(coef) z p age ph.karno tt(ph.karno) sex Likelihood ratio test=23.7 on 4 df, p=9.24e-05 n= 227, number of events= 164 (1 observation deleted due to missingness) Both the coefficients for ph.karno and tt(ph.karno) are statistically significant, implying that the effect of ph.karno varies with time. The time-varying effect of ph.karno can be written as β(t) = log(t + 20). We can add a line to the cox.zph plot of the time-varying effect of ph.karno on > zph.tt <- cox.zph(fit2, transform=function(t) log(t+20)) > plot(zph.tt[2]) > abline(coef(fit.tt)[2:3], col=2) The result is shown in Figure 3. The slope of the red line is 0.015, which is significantly different from the horizontal line (slope=0). The black line shows the time-varying coefficient for the variable ph.karno. Investigating time-varying coefficient with timereg package The timecox() function shipped with the timereg package (16) is able to fit a Cox model with both time-fixed and time-varying coefficients. In this case the time-varying effect is tested by resampling method (17). Specification of the models is similar to the way it is done in the survival package. > library(timereg) > fit.out <- timecox(surv(time,status)~ age+sex+ph.karno, data=lung,n.sim=500, max.time=700) Cox regression model is fit similarly as in the survival package with the only difference that resampling methods are used for the statistical inference and therefore, the number of simulations has to be specified (n.sim=500). The max.time argument specifies the end of observation period where estimates are computed. The returned results are shown below: > summary(fit.out) Multiplicative Hazard Model Test for nonparametric terms Test for non-significant effects Supremum-test p-value H_0:B(t)=0 of significance (Intercept) age sex ph.karno
8 Page 8 of 10 Zhang et al. -varying covariate and coefficient in survival analysis (Intercept) age Cumulative coefficients Cumulative coefficients sex ph.karno Cumulative coefficients Cumulative coefficients Figure 4 Estimated cumulative coefficients with 95% pointwise confidence intervals for intercept, age, sex and ph.karno. Test for time invariant effects Kolmogorov-Smirnov test p-value H_0: constant effect (Intercept) age sex ph.karno that the coefficients under test are not significantly different from 0), which shows that both sex and ph.karno have significant effect on survival outcome (P=0.002 and <0.001). The second table shows the test for the time invariant effect. Both the Kolmogorov-Smirnov test and the Cramer von Mises test are used for testing time invariant effects. It appears that the effect of ph.karno is not time-fixed. The effects of all variables over time are visualized in Figure 4. Cramer von Mises test p-value H_0: constant effect > par(mfrow=c(2,2)) (Intercept) 1.69e > plot(fit.out) age 9.82e sex 6.45e ph.karno 9.55e Call: timecox(formula = Surv(time, status) ~ age + sex + ph.karno, data = lung, max.time = 700, n.sim = 500) The first table of the output shows the results of the test for non-significant effect (e.g., the null hypothesis states It is noted from figure 4 that the effect of ph.karno is steep at the beginning and then flattens out after approximately 180. The variable age has no significant effect because the confidence interval intersects with the null effect reference line. The variable sex has significant effect but the null hypothesis of time invariance effect cannot be rejected. Therefore, we will proceed to set sex and age as time-fixed effect variables by fixing them with the const() function.
9 Annals of Translational Medicine, Vol 6, No 7 April 2018 Page 9 of 10 > fit.const <- timecox(surv(time,status)~ const(age)+const(sex)+ph.karno, data=lung,n.sim=500, max.time=700) > coef(fit.const) Coef. SE Robust SE z P-val lower2.5% upper97.5% const(age) const(sex) The const() function applied to the covariates age and sex specifies them to have constant effects. The age and sex have constant effects of and , respectively. Discussion When time-varying covariates or coefficients are present, an analyst should consider taking them into account in survival modeling in order to improve the estimation. In this paper, we presented some ways to do this using the R-program. -varying covariate was handled with a time-dependent Cox model and time-varying coefficient was described using a step function and a continuous function. In this article, we only presented some methods dealing with time-varying covariates or coefficients, but other approaches are available. Sometimes the model fit may also be improved by using derived variables from longitudinal measurements. For example, averages of the most recent and all the previous measurements may be used to better describe the cumulative nature of the time-varying covariate or differences of the latest two measurements to model the effects of changes (18). Also the standard deviation of the longitudinal measurements (19) and lagged observations (20) has been used. With internal time-varying covariates, one could also consider using joint modeling of longitudinal and survival data (9) which was not presented in this article. The idea is to assign a model for a continuously changing covariate which is measured longitudinally in time and possibly with error. This longitudinal model is related to survival times by modeling the joint distribution of longitudinal and survival data. Recent developments and issues in this topic are considered by, e.g., Hickey et al. (21). Acknowledgements Funding: The study was supported by Zhejiang Provincial Natural Science foundation of China (LGF18H150005). Footnote Conflicts of Interest: The authors have no conflicts of interest to declare. References 1. Zhang Z, Smischney NJ, Zhang H, et al. AME evidence series 001-The Society for Translational Medicine: clinical practice guidelines for diagnosis and early identification of sepsis in the hospital. J Thorac Dis 2016;8: Pinsky PF, Zhu CS, Kramer BS. Lung cancer risk by years since quitting in 30+ pack year smokers. J Med Screen 2015;22: Collett D. Modelling Survival Data in Medical Research. 3rd ed. Boca Raton: Chapman and Hall/CRC, Adeleke KA, Abiodun AA, Ipinyomi RA. Semi-parametric non-proportional hazard model with time varying covariate. Womens Health Journal 2015;14: Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Data. Hoboken, NJ, USA: John Wiley & Sons, Inc., Cox DR. Regression Models and Life-Tables. J R Statist Soc B 1972;34: Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. New York, NY: Springer New York, R Core Team. R: A language and environment for statistical computing [Internet] [cited 2018 Jan 8]. Available online: 9. Henderson R, Diggle P, Dobson A. Joint modelling of longitudinal measurements and event time data. Biostatistics 2000;1: Thomas L, Reyes EM. Tutorial: Survival Estimation for Cox Regression Models with -Varying Coefficients
10 Page 10 of 10 Zhang et al. -varying covariate and coefficient in survival analysis Using SASand R. Journal of Statistical Software 2014;61: Moriña D, Navarro A. Competing risks simulation with the survsim R package. Communications in Statistics - Simulation and Computation 2016;8: Austin PC. Generating survival times to simulate Cox proportional hazards models with time-varying covariates. Stat Med 2012;31: Andersen PK, Perme MP. Pseudo-observations in survival analysis. Stat Methods Med Res 2010;19: Therneau TM. A Package for Survival Analysis in S. version CRAN.R-project.orgpackagesurvival Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81: Martinussen T, Scheike TH. Dynamic Regression Models for Survival Data. Springer-Verlag New York, Tian L, Zucker D, Wei LJ. On the Cox Model With -Varying Regression Coefficients. Harvard University Biostatistics Working Paper 2005;100: Reinikainen J, Laatikainen T, Karvanen J, et al. Lifetime cumulative risk factors predict cardiovascular disease mortality in a 50-year follow-up study in Finland. Int J Epidemiol 2015;44: Muntner P, Shimbo D, Tonelli M, et al. The relationship between visit-to-visit variability in systolic blood pressure and all-cause mortality in the general population: findings from NHANES III, 1988 to Hypertension 2011;57: Hanson TE, Branscum AJ, Johnson WO. Predictive comparison of joint longitudinal-survival modeling: a case study illustrating competing approaches. Lifetime Data Anal 2011;17: Hickey GL, Philipson P, Jorgensen A, et al. Joint modelling of time-to-event and multivariate longitudinal outcomes: recent developments and issues. BMC Med Res Methodol 2016;16:117. Cite this article as: Zhang Z, Reinikainen J, Adeleke KA, Pieterse ME, Groothuis-Oudshoorn CG. -varying covariates and coefficients in Cox regression models. Ann Transl Med 2018;6(7):121. doi: /atm
Lecture 7 Time-dependent Covariates in Cox Regression
Lecture 7 Time-dependent Covariates in Cox Regression So far, we ve been considering the following Cox PH model: λ(t Z) = λ 0 (t) exp(β Z) = λ 0 (t) exp( β j Z j ) where β j is the parameter for the the
More informationTime-dependent coefficients
Time-dependent coefficients Patrick Breheny December 1 Patrick Breheny Survival Data Analysis (BIOS 7210) 1/20 Introduction As we discussed previously, stratification allows one to handle variables that
More informationThe coxvc_1-1-1 package
Appendix A The coxvc_1-1-1 package A.1 Introduction The coxvc_1-1-1 package is a set of functions for survival analysis that run under R2.1.1 [81]. This package contains a set of routines to fit Cox models
More informationSurvival analysis in R
Survival analysis in R Niels Richard Hansen This note describes a few elementary aspects of practical analysis of survival data in R. For further information we refer to the book Introductory Statistics
More informationMAS3301 / MAS8311 Biostatistics Part II: Survival
MAS3301 / MAS8311 Biostatistics Part II: Survival M. Farrow School of Mathematics and Statistics Newcastle University Semester 2, 2009-10 1 13 The Cox proportional hazards model 13.1 Introduction In the
More informationSurvival Regression Models
Survival Regression Models David M. Rocke May 18, 2017 David M. Rocke Survival Regression Models May 18, 2017 1 / 32 Background on the Proportional Hazards Model The exponential distribution has constant
More informationLecture 12. Multivariate Survival Data Statistics Survival Analysis. Presented March 8, 2016
Statistics 255 - Survival Analysis Presented March 8, 2016 Dan Gillen Department of Statistics University of California, Irvine 12.1 Examples Clustered or correlated survival times Disease onset in family
More informationExtensions of Cox Model for Non-Proportional Hazards Purpose
PhUSE 2013 Paper SP07 Extensions of Cox Model for Non-Proportional Hazards Purpose Jadwiga Borucka, PAREXEL, Warsaw, Poland ABSTRACT Cox proportional hazard model is one of the most common methods used
More informationResiduals and regression diagnostics: focusing on logistic regression
Big-data Clinical Trial Column Page of 8 Residuals and regression diagnostics: focusing on logistic regression Zhongheng Zhang Department of Critical Care Medicine, Jinhua Municipal Central Hospital, Jinhua
More informationFULL LIKELIHOOD INFERENCES IN THE COX MODEL
October 20, 2007 FULL LIKELIHOOD INFERENCES IN THE COX MODEL BY JIAN-JIAN REN 1 AND MAI ZHOU 2 University of Central Florida and University of Kentucky Abstract We use the empirical likelihood approach
More informationREGRESSION ANALYSIS FOR TIME-TO-EVENT DATA THE PROPORTIONAL HAZARDS (COX) MODEL ST520
REGRESSION ANALYSIS FOR TIME-TO-EVENT DATA THE PROPORTIONAL HAZARDS (COX) MODEL ST520 Department of Statistics North Carolina State University Presented by: Butch Tsiatis, Department of Statistics, NCSU
More information1 Introduction. 2 Residuals in PH model
Supplementary Material for Diagnostic Plotting Methods for Proportional Hazards Models With Time-dependent Covariates or Time-varying Regression Coefficients BY QIQING YU, JUNYI DONG Department of Mathematical
More informationExtensions of Cox Model for Non-Proportional Hazards Purpose
PhUSE Annual Conference 2013 Paper SP07 Extensions of Cox Model for Non-Proportional Hazards Purpose Author: Jadwiga Borucka PAREXEL, Warsaw, Poland Brussels 13 th - 16 th October 2013 Presentation Plan
More informationChecking model assumptions with regression diagnostics
@graemeleehickey www.glhickey.com graeme.hickey@liverpool.ac.uk Checking model assumptions with regression diagnostics Graeme L. Hickey University of Liverpool Conflicts of interest None Assistant Editor
More informationLecture 8 Stat D. Gillen
Statistics 255 - Survival Analysis Presented February 23, 2016 Dan Gillen Department of Statistics University of California, Irvine 8.1 Example of two ways to stratify Suppose a confounder C has 3 levels
More informationSurvival Analysis. 732G34 Statistisk analys av komplexa data. Krzysztof Bartoszek
Survival Analysis 732G34 Statistisk analys av komplexa data Krzysztof Bartoszek (krzysztof.bartoszek@liu.se) 10, 11 I 2018 Department of Computer and Information Science Linköping University Survival analysis
More informationIn contrast, parametric techniques (fitting exponential or Weibull, for example) are more focussed, can handle general covariates, but require
Chapter 5 modelling Semi parametric We have considered parametric and nonparametric techniques for comparing survival distributions between different treatment groups. Nonparametric techniques, such as
More informationUNIVERSITY OF CALIFORNIA, SAN DIEGO
UNIVERSITY OF CALIFORNIA, SAN DIEGO Estimation of the primary hazard ratio in the presence of a secondary covariate with non-proportional hazards An undergraduate honors thesis submitted to the Department
More informationPackage complex.surv.dat.sim
Package complex.surv.dat.sim February 15, 2013 Type Package Title Simulation of complex survival data Version 1.1.1 Date 2012-07-09 Encoding latin1 Author David Moriña, Centre Tecnològic de Nutrició i
More informationGoodness-Of-Fit for Cox s Regression Model. Extensions of Cox s Regression Model. Survival Analysis Fall 2004, Copenhagen
Outline Cox s proportional hazards model. Goodness-of-fit tools More flexible models R-package timereg Forthcoming book, Martinussen and Scheike. 2/38 University of Copenhagen http://www.biostat.ku.dk
More informationMultivariate Survival Analysis
Multivariate Survival Analysis Previously we have assumed that either (X i, δ i ) or (X i, δ i, Z i ), i = 1,..., n, are i.i.d.. This may not always be the case. Multivariate survival data can arise in
More informationLecture 9. Statistics Survival Analysis. Presented February 23, Dan Gillen Department of Statistics University of California, Irvine
Statistics 255 - Survival Analysis Presented February 23, 2016 Dan Gillen Department of Statistics University of California, Irvine 9.1 Survival analysis involves subjects moving through time Hazard may
More informationYou know I m not goin diss you on the internet Cause my mama taught me better than that I m a survivor (What?) I m not goin give up (What?
You know I m not goin diss you on the internet Cause my mama taught me better than that I m a survivor (What?) I m not goin give up (What?) I m not goin stop (What?) I m goin work harder (What?) Sir David
More informationSurvival analysis in R
Survival analysis in R Niels Richard Hansen This note describes a few elementary aspects of practical analysis of survival data in R. For further information we refer to the book Introductory Statistics
More informationLongitudinal + Reliability = Joint Modeling
Longitudinal + Reliability = Joint Modeling Carles Serrat Institute of Statistics and Mathematics Applied to Building CYTED-HAROSA International Workshop November 21-22, 2013 Barcelona Mainly from Rizopoulos,
More informationNeural networks: further insights into error function, generalized weights and others
Big-data Clinical Trial Column Page 1 of 6 Neural networks: further insights into error function, generalized weights and others Zhongheng Zhang Department of Critical Care Medicine, Jinhua Municipal Central
More informationOn the Breslow estimator
Lifetime Data Anal (27) 13:471 48 DOI 1.17/s1985-7-948-y On the Breslow estimator D. Y. Lin Received: 5 April 27 / Accepted: 16 July 27 / Published online: 2 September 27 Springer Science+Business Media,
More informationJournal of Statistical Software
JSS Journal of Statistical Software January 2011, Volume 38, Issue 2. http://www.jstatsoft.org/ Analyzing Competing Risk Data Using the R timereg Package Thomas H. Scheike University of Copenhagen Mei-Jie
More informationβ j = coefficient of x j in the model; β = ( β1, β2,
Regression Modeling of Survival Time Data Why regression models? Groups similar except for the treatment under study use the nonparametric methods discussed earlier. Groups differ in variables (covariates)
More informationMultivariable Fractional Polynomials
Multivariable Fractional Polynomials Axel Benner September 7, 2015 Contents 1 Introduction 1 2 Inventory of functions 1 3 Usage in R 2 3.1 Model selection........................................ 3 4 Example
More informationCox s proportional hazards/regression model - model assessment
Cox s proportional hazards/regression model - model assessment Rasmus Waagepetersen September 27, 2017 Topics: Plots based on estimated cumulative hazards Cox-Snell residuals: overall check of fit Martingale
More informationA Regression Model For Recurrent Events With Distribution Free Correlation Structure
A Regression Model For Recurrent Events With Distribution Free Correlation Structure J. Pénichoux(1), A. Latouche(2), T. Moreau(1) (1) INSERM U780 (2) Université de Versailles, EA2506 ISCB - 2009 - Prague
More informationRelative-risk regression and model diagnostics. 16 November, 2015
Relative-risk regression and model diagnostics 16 November, 2015 Relative risk regression More general multiplicative intensity model: Intensity for individual i at time t is i(t) =Y i (t)r(x i, ; t) 0
More informationMultivariable Fractional Polynomials
Multivariable Fractional Polynomials Axel Benner May 17, 2007 Contents 1 Introduction 1 2 Inventory of functions 1 3 Usage in R 2 3.1 Model selection........................................ 3 4 Example
More informationLecture 7. Proportional Hazards Model - Handling Ties and Survival Estimation Statistics Survival Analysis. Presented February 4, 2016
Proportional Hazards Model - Handling Ties and Survival Estimation Statistics 255 - Survival Analysis Presented February 4, 2016 likelihood - Discrete Dan Gillen Department of Statistics University of
More informationADVANCED STATISTICAL ANALYSIS OF EPIDEMIOLOGICAL STUDIES. Cox s regression analysis Time dependent explanatory variables
ADVANCED STATISTICAL ANALYSIS OF EPIDEMIOLOGICAL STUDIES Cox s regression analysis Time dependent explanatory variables Henrik Ravn Bandim Health Project, Statens Serum Institut 4 November 2011 1 / 53
More informationMultistate models and recurrent event models
Multistate models Multistate models and recurrent event models Patrick Breheny December 10 Patrick Breheny Survival Data Analysis (BIOS 7210) 1/22 Introduction Multistate models In this final lecture,
More informationPackage crrsc. R topics documented: February 19, 2015
Package crrsc February 19, 2015 Title Competing risks regression for Stratified and Clustered data Version 1.1 Author Bingqing Zhou and Aurelien Latouche Extension of cmprsk to Stratified and Clustered
More informationResiduals and model diagnostics
Residuals and model diagnostics Patrick Breheny November 10 Patrick Breheny Survival Data Analysis (BIOS 7210) 1/42 Introduction Residuals Many assumptions go into regression models, and the Cox proportional
More informationOn a connection between the Bradley-Terry model and the Cox proportional hazards model
On a connection between the Bradley-Terry model and the Cox proportional hazards model Yuhua Su and Mai Zhou Department of Statistics University of Kentucky Lexington, KY 40506-0027, U.S.A. SUMMARY This
More informationPrerequisite: STATS 7 or STATS 8 or AP90 or (STATS 120A and STATS 120B and STATS 120C). AP90 with a minimum score of 3
University of California, Irvine 2017-2018 1 Statistics (STATS) Courses STATS 5. Seminar in Data Science. 1 Unit. An introduction to the field of Data Science; intended for entering freshman and transfers.
More informationLecture 10. Diagnostics. Statistics Survival Analysis. Presented March 1, 2016
Statistics 255 - Survival Analysis Presented March 1, 2016 Dan Gillen Department of Statistics University of California, Irvine 10.1 Are model assumptions correct? Is the proportional hazards assumption
More informationTMA 4275 Lifetime Analysis June 2004 Solution
TMA 4275 Lifetime Analysis June 2004 Solution Problem 1 a) Observation of the outcome is censored, if the time of the outcome is not known exactly and only the last time when it was observed being intact,
More informationPackage survsim. R topics documented: October 16, 2015
Package survsim October 16, 2015 Type Package Title Simulation of Simple and Comple Survival Data Version 1.1.4 Date 2015-10-07 Encoding UTF-8 Author David Moriña (Centre for Research in Environmental
More informationMultistate models and recurrent event models
and recurrent event models Patrick Breheny December 6 Patrick Breheny University of Iowa Survival Data Analysis (BIOS:7210) 1 / 22 Introduction In this final lecture, we will briefly look at two other
More informationSTAT331. Cox s Proportional Hazards Model
STAT331 Cox s Proportional Hazards Model In this unit we introduce Cox s proportional hazards (Cox s PH) model, give a heuristic development of the partial likelihood function, and discuss adaptations
More informationPart [1.0] Measures of Classification Accuracy for the Prediction of Survival Times
Part [1.0] Measures of Classification Accuracy for the Prediction of Survival Times Patrick J. Heagerty PhD Department of Biostatistics University of Washington 1 Biomarkers Review: Cox Regression Model
More informationApplied Survival Analysis Lab 10: Analysis of multiple failures
Applied Survival Analysis Lab 10: Analysis of multiple failures We will analyze the bladder data set (Wei et al., 1989). A listing of the dataset is given below: list if id in 1/9 +---------------------------------------------------------+
More informationMulti-state Models: An Overview
Multi-state Models: An Overview Andrew Titman Lancaster University 14 April 2016 Overview Introduction to multi-state modelling Examples of applications Continuously observed processes Intermittently observed
More informationTime-dependent covariates
Time-dependent covariates Rasmus Waagepetersen November 5, 2018 1 / 10 Time-dependent covariates Our excursion into the realm of counting process and martingales showed that it poses no problems to introduce
More informationA fast routine for fitting Cox models with time varying effects
Chapter 3 A fast routine for fitting Cox models with time varying effects Abstract The S-plus and R statistical packages have implemented a counting process setup to estimate Cox models with time varying
More informationChapter 4 Regression Models
23.August 2010 Chapter 4 Regression Models The target variable T denotes failure time We let x = (x (1),..., x (m) ) represent a vector of available covariates. Also called regression variables, regressors,
More informationOn a connection between the Bradley Terry model and the Cox proportional hazards model
Statistics & Probability Letters 76 (2006) 698 702 www.elsevier.com/locate/stapro On a connection between the Bradley Terry model and the Cox proportional hazards model Yuhua Su, Mai Zhou Department of
More informationJoint Modeling of Longitudinal Item Response Data and Survival
Joint Modeling of Longitudinal Item Response Data and Survival Jean-Paul Fox University of Twente Department of Research Methodology, Measurement and Data Analysis Faculty of Behavioural Sciences Enschede,
More informationSTAT 526 Spring Final Exam. Thursday May 5, 2011
STAT 526 Spring 2011 Final Exam Thursday May 5, 2011 Time: 2 hours Name (please print): Show all your work and calculations. Partial credit will be given for work that is partially correct. Points will
More informationPower and Sample Size Calculations with the Additive Hazards Model
Journal of Data Science 10(2012), 143-155 Power and Sample Size Calculations with the Additive Hazards Model Ling Chen, Chengjie Xiong, J. Philip Miller and Feng Gao Washington University School of Medicine
More informationCompare Predicted Counts between Groups of Zero Truncated Poisson Regression Model based on Recycled Predictions Method
Compare Predicted Counts between Groups of Zero Truncated Poisson Regression Model based on Recycled Predictions Method Yan Wang 1, Michael Ong 2, Honghu Liu 1,2,3 1 Department of Biostatistics, UCLA School
More informationRonald Christensen. University of New Mexico. Albuquerque, New Mexico. Wesley Johnson. University of California, Irvine. Irvine, California
Texts in Statistical Science Bayesian Ideas and Data Analysis An Introduction for Scientists and Statisticians Ronald Christensen University of New Mexico Albuquerque, New Mexico Wesley Johnson University
More information4. Comparison of Two (K) Samples
4. Comparison of Two (K) Samples K=2 Problem: compare the survival distributions between two groups. E: comparing treatments on patients with a particular disease. Z: Treatment indicator, i.e. Z = 1 for
More informationPhilosophy and Features of the mstate package
Introduction Mathematical theory Practice Discussion Philosophy and Features of the mstate package Liesbeth de Wreede, Hein Putter Department of Medical Statistics and Bioinformatics Leiden University
More informationLongitudinal Modeling with Logistic Regression
Newsom 1 Longitudinal Modeling with Logistic Regression Longitudinal designs involve repeated measurements of the same individuals over time There are two general classes of analyses that correspond to
More informationCHL 5225 H Crossover Trials. CHL 5225 H Crossover Trials
CHL 55 H Crossover Trials The Two-sequence, Two-Treatment, Two-period Crossover Trial Definition A trial in which patients are randomly allocated to one of two sequences of treatments (either 1 then, or
More informationA COMPARISON OF POISSON AND BINOMIAL EMPIRICAL LIKELIHOOD Mai Zhou and Hui Fang University of Kentucky
A COMPARISON OF POISSON AND BINOMIAL EMPIRICAL LIKELIHOOD Mai Zhou and Hui Fang University of Kentucky Empirical likelihood with right censored data were studied by Thomas and Grunkmier (1975), Li (1995),
More informationST745: Survival Analysis: Cox-PH!
ST745: Survival Analysis: Cox-PH! Eric B. Laber Department of Statistics, North Carolina State University April 20, 2015 Rien n est plus dangereux qu une idee, quand on n a qu une idee. (Nothing is more
More informationIntroduction to mtm: An R Package for Marginalized Transition Models
Introduction to mtm: An R Package for Marginalized Transition Models Bryan A. Comstock and Patrick J. Heagerty Department of Biostatistics University of Washington 1 Introduction Marginalized transition
More informationThe Proportional Hazard Model and the Modelling of Recurrent Failure Data: Analysis of a Disconnector Population in Sweden. Sweden
PS1 Life Cycle Asset Management The Proportional Hazard Model and the Modelling of Recurrent Failure Data: Analysis of a Disconnector Population in Sweden J. H. Jürgensen 1, A.L. Brodersson 2, P. Hilber
More informationStatistics in medicine
Statistics in medicine Lecture 4: and multivariable regression Fatma Shebl, MD, MS, MPH, PhD Assistant Professor Chronic Disease Epidemiology Department Yale School of Public Health Fatma.shebl@yale.edu
More informationDescription Syntax for predict Menu for predict Options for predict Remarks and examples Methods and formulas References Also see
Title stata.com stcrreg postestimation Postestimation tools for stcrreg Description Syntax for predict Menu for predict Options for predict Remarks and examples Methods and formulas References Also see
More informationPackage CoxRidge. February 27, 2015
Type Package Title Cox Models with Dynamic Ridge Penalties Version 0.9.2 Date 2015-02-12 Package CoxRidge February 27, 2015 Author Aris Perperoglou Maintainer Aris Perperoglou
More informationAccelerated Failure Time Models: A Review
International Journal of Performability Engineering, Vol. 10, No. 01, 2014, pp.23-29. RAMS Consultants Printed in India Accelerated Failure Time Models: A Review JEAN-FRANÇOIS DUPUY * IRMAR/INSA of Rennes,
More informationLecture 11. Interval Censored and. Discrete-Time Data. Statistics Survival Analysis. Presented March 3, 2016
Statistics 255 - Survival Analysis Presented March 3, 2016 Motivating Dan Gillen Department of Statistics University of California, Irvine 11.1 First question: Are the data truly discrete? : Number of
More informationLecture 5 Models and methods for recurrent event data
Lecture 5 Models and methods for recurrent event data Recurrent and multiple events are commonly encountered in longitudinal studies. In this chapter we consider ordered recurrent and multiple events.
More informationPart III Measures of Classification Accuracy for the Prediction of Survival Times
Part III Measures of Classification Accuracy for the Prediction of Survival Times Patrick J Heagerty PhD Department of Biostatistics University of Washington 102 ISCB 2010 Session Three Outline Examples
More informationPart IV Extensions: Competing Risks Endpoints and Non-Parametric AUC(t) Estimation
Part IV Extensions: Competing Risks Endpoints and Non-Parametric AUC(t) Estimation Patrick J. Heagerty PhD Department of Biostatistics University of Washington 166 ISCB 2010 Session Four Outline Examples
More informationLogistic regression model for survival time analysis using time-varying coefficients
Logistic regression model for survival time analysis using time-varying coefficients Accepted in American Journal of Mathematical and Management Sciences, 2016 Kenichi SATOH ksatoh@hiroshima-u.ac.jp Research
More informationFirst Aid Kit for Survival. Hypoxia cohort. Goal. DFS=Clinical + Marker 1/21/2015. Two analyses to exemplify some concepts of survival techniques
First Aid Kit for Survival Melania Pintilie pintilie@uhnres.utoronto.ca Two analyses to exemplify some concepts of survival techniques Checking linearity Checking proportionality of hazards Predicted curves:
More informationUniversity of California, Berkeley
University of California, Berkeley U.C. Berkeley Division of Biostatistics Working Paper Series Year 24 Paper 153 A Note on Empirical Likelihood Inference of Residual Life Regression Ying Qing Chen Yichuan
More informationApproximation of Survival Function by Taylor Series for General Partly Interval Censored Data
Malaysian Journal of Mathematical Sciences 11(3): 33 315 (217) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES Journal homepage: http://einspem.upm.edu.my/journal Approximation of Survival Function by Taylor
More informationMulti-state models: prediction
Department of Medical Statistics and Bioinformatics Leiden University Medical Center Course on advanced survival analysis, Copenhagen Outline Prediction Theory Aalen-Johansen Computational aspects Applications
More informationRobust estimates of state occupancy and transition probabilities for Non-Markov multi-state models
Robust estimates of state occupancy and transition probabilities for Non-Markov multi-state models 26 March 2014 Overview Continuously observed data Three-state illness-death General robust estimator Interval
More informationSSUI: Presentation Hints 2 My Perspective Software Examples Reliability Areas that need work
SSUI: Presentation Hints 1 Comparing Marginal and Random Eects (Frailty) Models Terry M. Therneau Mayo Clinic April 1998 SSUI: Presentation Hints 2 My Perspective Software Examples Reliability Areas that
More informationTypical Survival Data Arising From a Clinical Trial. Censoring. The Survivor Function. Mathematical Definitions Introduction
Outline CHL 5225H Advanced Statistical Methods for Clinical Trials: Survival Analysis Prof. Kevin E. Thorpe Defining Survival Data Mathematical Definitions Non-parametric Estimates of Survival Comparing
More informationQuantile Regression for Residual Life and Empirical Likelihood
Quantile Regression for Residual Life and Empirical Likelihood Mai Zhou email: mai@ms.uky.edu Department of Statistics, University of Kentucky, Lexington, KY 40506-0027, USA Jong-Hyeon Jeong email: jeong@nsabp.pitt.edu
More informationA TWO-STAGE LINEAR MIXED-EFFECTS/COX MODEL FOR LONGITUDINAL DATA WITH MEASUREMENT ERROR AND SURVIVAL
A TWO-STAGE LINEAR MIXED-EFFECTS/COX MODEL FOR LONGITUDINAL DATA WITH MEASUREMENT ERROR AND SURVIVAL Christopher H. Morrell, Loyola College in Maryland, and Larry J. Brant, NIA Christopher H. Morrell,
More informationDynamic Prediction of Disease Progression Using Longitudinal Biomarker Data
Dynamic Prediction of Disease Progression Using Longitudinal Biomarker Data Xuelin Huang Department of Biostatistics M. D. Anderson Cancer Center The University of Texas Joint Work with Jing Ning, Sangbum
More informationFULL LIKELIHOOD INFERENCES IN THE COX MODEL: AN EMPIRICAL LIKELIHOOD APPROACH
FULL LIKELIHOOD INFERENCES IN THE COX MODEL: AN EMPIRICAL LIKELIHOOD APPROACH Jian-Jian Ren 1 and Mai Zhou 2 University of Central Florida and University of Kentucky Abstract: For the regression parameter
More informationSimulation-based robust IV inference for lifetime data
Simulation-based robust IV inference for lifetime data Anand Acharya 1 Lynda Khalaf 1 Marcel Voia 1 Myra Yazbeck 2 David Wensley 3 1 Department of Economics Carleton University 2 Department of Economics
More informationBIOL 51A - Biostatistics 1 1. Lecture 1: Intro to Biostatistics. Smoking: hazardous? FEV (l) Smoke
BIOL 51A - Biostatistics 1 1 Lecture 1: Intro to Biostatistics Smoking: hazardous? FEV (l) 1 2 3 4 5 No Yes Smoke BIOL 51A - Biostatistics 1 2 Box Plot a.k.a box-and-whisker diagram or candlestick chart
More informationThe nltm Package. July 24, 2006
The nltm Package July 24, 2006 Version 1.2 Date 2006-07-17 Title Non-linear Transformation Models Author Gilda Garibotti, Alexander Tsodikov Maintainer Gilda Garibotti Depends
More informationOther Survival Models. (1) Non-PH models. We briefly discussed the non-proportional hazards (non-ph) model
Other Survival Models (1) Non-PH models We briefly discussed the non-proportional hazards (non-ph) model λ(t Z) = λ 0 (t) exp{β(t) Z}, where β(t) can be estimated by: piecewise constants (recall how);
More informationMüller: Goodness-of-fit criteria for survival data
Müller: Goodness-of-fit criteria for survival data Sonderforschungsbereich 386, Paper 382 (2004) Online unter: http://epub.ub.uni-muenchen.de/ Projektpartner Goodness of fit criteria for survival data
More informationA comparison study of the nonparametric tests based on the empirical distributions
통계연구 (2015), 제 20 권제 3 호, 1-12 A comparison study of the nonparametric tests based on the empirical distributions Hyo-Il Park 1) Abstract In this study, we propose a nonparametric test based on the empirical
More informationPackage idmtpreg. February 27, 2018
Type Package Package idmtpreg February 27, 2018 Title Regression Model for Progressive Illness Death Data Version 1.1 Date 2018-02-23 Author Leyla Azarang and Manuel Oviedo de la Fuente Maintainer Leyla
More informationSemiparametric Regression
Semiparametric Regression Patrick Breheny October 22 Patrick Breheny Survival Data Analysis (BIOS 7210) 1/23 Introduction Over the past few weeks, we ve introduced a variety of regression models under
More informationSTA6938-Logistic Regression Model
Dr. Ying Zhang STA6938-Logistic Regression Model Topic 6-Logistic Regression for Case-Control Studies Outlines: 1. Biomedical Designs 2. Logistic Regression Models for Case-Control Studies 3. Logistic
More informationAnalysis of Time-to-Event Data: Chapter 6 - Regression diagnostics
Analysis of Time-to-Event Data: Chapter 6 - Regression diagnostics Steffen Unkel Department of Medical Statistics University Medical Center Göttingen, Germany Winter term 2018/19 1/25 Residuals for the
More informationPackage threg. August 10, 2015
Package threg August 10, 2015 Title Threshold Regression Version 1.0.3 Date 2015-08-10 Author Tao Xiao Maintainer Tao Xiao Depends R (>= 2.10), survival, Formula Fit a threshold regression
More informationReduced-rank hazard regression
Chapter 2 Reduced-rank hazard regression Abstract The Cox proportional hazards model is the most common method to analyze survival data. However, the proportional hazards assumption might not hold. The
More informationBIAS OF MAXIMUM-LIKELIHOOD ESTIMATES IN LOGISTIC AND COX REGRESSION MODELS: A COMPARATIVE SIMULATION STUDY
BIAS OF MAXIMUM-LIKELIHOOD ESTIMATES IN LOGISTIC AND COX REGRESSION MODELS: A COMPARATIVE SIMULATION STUDY Ingo Langner 1, Ralf Bender 2, Rebecca Lenz-Tönjes 1, Helmut Küchenhoff 2, Maria Blettner 2 1
More informationTutorial 6: Tutorial on Translating between GLIMMPSE Power Analysis and Data Analysis. Acknowledgements:
Tutorial 6: Tutorial on Translating between GLIMMPSE Power Analysis and Data Analysis Anna E. Barón, Keith E. Muller, Sarah M. Kreidler, and Deborah H. Glueck Acknowledgements: The project was supported
More information